100 REPORTS ON THE STATE OF SCIENCE.—1917. viy ee y 
of these observations is to show that calcium metasilicate is’ present 
both in the crystalline and the gelatinous forms, but that the tricalcium 
aluminate is entirely crystalline. The effect of stains varies considerably 
with the conditions of the experiment, and the present writer has been 
in many instances unable to confirm the observations of Keisermann. 
It should be added that so experienced an observer as Le Chatelier still 
considers that the process of setting of hydraulic cements is entirely one 
of crystallisation, and denies the influence of colloidal substances.1* 
In repeating experiments on this point, it is important to remember 
that the process is not necessarily the same when a small quantity of 
cement is mixed with a relatively large proportion of water on a microscope 
slide and when, as in the practical use of cement or the making of test 
briquettes, the cement and water are mixed to a stiff paste; and neglect 
of this fact, no doubt, accounts for some of the divergences between the 
conclusions of different investigators. The influence of the size of grain 
was pointed out by Wetzel in a criticism of the work of Keisermann.'4 
An explanation of the hardening of cement in air, based on the assumption 
of the presence of colloids, which then absorb water vapour from the 
air, was given by L. Jesser.!® 
Experiments in which stains were used led F. Blumenthal +® to the 
conclusion that crystalline monocalcium silicate was among the first 
products of hydration, together with the aluminate, and that the formation 
of the gelatinous silicate took place subsequently. The setting of the 
cement was then due to crystallisation alone, the later hardening being 
due to the binding together of the crystals by means of a gel, and the 
filling of the pores in the same manner. 
The chemistry of Portland cement was set on an entirely new basis 
by the splendid investigations of the lime-silica-alumina system in 
the Geophysical Laboratory of the Carnegie Institution, Washington. 
As the result of these investigations, the constitution of cement clinker 
became known in detail, and it was then possible to consider the process 
of hydration as one of a definite series of chemical reactions. The study 
of the setting and hardening processes was taken up by the staff of the 
U.S. Bureau of Standards, whose results have been published in the 
form of monographs.1? The general conclusions may be summarised 
as follows. 
Tricalcium aluminate sets so rapidly that it is practically impossible 
to form it into test pieces, but it develops little strength. The hydrated 
product is mostly crystalline. The aluminate 5Ca0,Al;0; also hydrates 
rapidly, forming an amorphous mass, which partly crystallises, the product 
being apparently the hydrate of tricalcium aluminate. Dicalcium silicate 
hydrates exceedingly slowly, and the product is entirely amorphous, 
except for the crystals of calcium hydroxide which make their appearance 
after a time. It is not believed that this amorphous or colloidal mass 
corresponds with any definite calcium silicate, but rather that it is an 
13 Private communication. 
14 Zentr. hydraul. Zemente, 1911, 2, 34. 
16 [bid., 51. ; 
16 Dissertation, Jena, 1912; Silikat Zeitschr., 1914, 2, 43. 
17 A, A. Klein and A. J. Phillips, U.S. Bureau of Standards, Technological Paper 43 
(1914); P. H. Bates and A. A. Klein, Technol. Paper 78 (1916). Seealso G. A. Rankin, 
J, Franklin Inst. 1916, 747. 
